METR Benchmark — leaderboard
Metric: 50% Time Horizon (hours). Source: metr.org. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Mythos Preview | 17.41 |
| 2 | Claude Opus 4.6 | 11.98 |
| 3 | Gemini 3.1 Pro (Preview) | 6.4 |
| 4 | GPT-5.2 | 5.87 |
| 5 | GPT-5.3 Codex | 5.83 |
| 6 | GPT-5.4 | 5.7 |
| 7 | Claude Opus 4.5 | 4.88 |
| 8 | Gemini 3 Pro | 3.74 |
| 9 | GPT-5.1 Codex Max | 3.73 |
| 10 | GPT-5 | 3.38 |
| 11 | O3 | 2 |
| 12 | Claude Opus 4 | 1.67 |
| 13 | Claude Opus 4.1 | 1.67 |
| 14 | Claude 3.7 Sonnet | 1.01 |
| 15 | O1 | 0.65 |
Interactive version: theaggregate.ai/benchmark?slug=metr-benchmark · How the rankings work · Data refreshed daily, snapshot 2026-07-22.